Stabilize and improve cars-rest perf suite 09/73209/18
authorJamo Luhrsen <jluhrsen@redhat.com>
Tue, 19 Jun 2018 21:57:43 +0000 (14:57 -0700)
committerJamo Luhrsen <jluhrsen@redhat.com>
Thu, 5 Jul 2018 20:41:54 +0000 (13:41 -0700)
commit967f534f762ce38f7cf2f1bb1139a8d9ab4205c0
treef47da5e851fcdb3bf65df37390400885ca68b77b
parent6d7719503cb948051d3cd9f84041f3409a4e13c8
Stabilize and improve cars-rest perf suite

- use a more strict prompt for checking when test tool is
  done. Some 500 error responses (which are ok when leaders
  are down) gave a > prompt which was being wrongly interpreted
  as the system prompt. now, making it look for ~]> instead.

- running the suites twice now, so we can see ask based and
  tell based protocols

- making the python client retry on 500 errors which can come
  when leaders are down and we have askTimeouts

- polling a little longer on some steps

Change-Id: I426f1b03f2fdd3258a39c03bd8e568f16beebf88
Signed-off-by: Jamo Luhrsen <jluhrsen@redhat.com>
csit/libraries/SSHKeywords.robot
csit/libraries/Utils.robot
csit/suites/controller/ThreeNodes_Datastore/puts_during_isolation.robot
csit/testplans/controller-rest-clust-cars-perf.txt
tools/odl-mdsal-clustering-tests/scripts/cluster_rest_script.py